Parallel FPGA technology mapping using multi-core architectures

We present a parallelization of a modern FPGA technology mapper suitable for use on a multi-core processor which exploits the inherent data parallelism of the algorithm. Our parallel mapper is deterministic regardless of the number of cores available which is important in an industrial context. Without any compromise in the quality of result, we show average speed-ups of 1.5X to 3.2X on a dual-core and quad-core multi-processor, respectively.

[1]  Bradford Nichols,et al.  Pthreads programming , 1996 .

[2]  Vamsi Boppana,et al.  A Parallel Algorithm for the Technology Mapping of LUT-Based FPGAs , 1996, Euro-Par, Vol. I.

[3]  James Reinders,et al.  Intel® threading building blocks , 2008 .

[4]  Robert K. Brayton,et al.  Improvements to Technology Mapping for LUT-Based FPGAs , 2007, IEEE Trans. Comput. Aided Des. Integr. Circuits Syst..

[5]  Jason Cong,et al.  DAOmap: a depth-optimal area optimization mapping algorithm for FPGA designs , 2004, IEEE/ACM International Conference on Computer Aided Design, 2004. ICCAD-2004..

[6]  Stephen Dean Brown,et al.  Heuristics for Area Minimization in LUT-Based FPGA Technology Mapping , 2006, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[7]  Martine D. F. Schlag,et al.  Parallel placement for field-programmable gate arrays , 2003, FPGA '03.

[8]  Jason Cong,et al.  FlowMap: an optimal technology mapping algorithm for delay optimization in lookup-table based FPGA designs , 1994, IEEE Trans. Comput. Aided Des. Integr. Circuits Syst..

[9]  Doris Chen,et al.  Parallelizing FPGA Technology Mapping Using Graphics Processing Units (GPUs) , 2010, 2010 International Conference on Field Programmable Logic and Applications.

[10]  Vaughn Betz,et al.  High-quality, deterministic parallel placement for FPGAs on commodity hardware , 2008, FPGA '08.